There has been a spike of burglary in the Bay Area, and we felt that security with facial recognition will help make things safer for homes and small businesses.

What it does

When there is someone nearby PiWatch detects the intruder with the pi motion sensor. It fires up its recognition program to check if the intruder is safe or not. With our machine learning software, the computer can recognise if the intruder is someone you deem safe or is unknown. It gives a percentage on how likely this person is a registered safe person. The more that our software sees known people the better it becomes at recognising who is safe and who is not. Notifications are sent if PiWatch identifies an unknown intruder. They receive an email notifying them that there is danger, letting you know you need to check your camera to make sure your home or property is safe!

How we built it

PiWatch software was built with python and RaspberryPi. We used machine learning for facial recognition, training it with pictures of people who are to be recognised as safe.

The website was the first framework on Figma. Then we moved to Replit using html and the CSS from Figma, the website was created.

Challenges we ran into

When creating the website there were some challenges with integrating the CSS code to make it look how we want. Another issue we ran into was the pi cam breaking. Thankfully we were able to get another working one and was able to continue with our project. It was also challenging learning how to use machine learning and training it. Collecting and providing the data for ML was very hard. In fact there were other ideas we wanted to do with ML, but we were sadly not able to do them because we lacked enough data for it nor had the time to collect it all.

Accomplishments that we're proud of

We are very proud of the website design. It looks very professional and clean and we are very happy with it. We are also happy that we were able to finish our project and have something to present.

What we learned

We learned that data collection and training for ML is very hard and a lot of data is needed for it to give good results. We also learned that Figma is a great tool for creating websites. This was our first time using Figma to create a website and we are very happy with it. Figma makes it so much easier to create clean and nice looking websites.

What's next for PiWatch

PiWatch will update our software where users are able to use it with security cameras. We also want to create an app that will notify users (instead of email) and to watch the security feed.

Built With

Share this project: